Hurricane-force winds gave at most a small decrease in next-summer resident bird counts on 2,025 eastern US survey routes in 1970-2024, and the decrease did not repeat across regions and periods

Question: do Atlantic hurricanes decrease the next breeding-season counts of resident birds on North American Breeding Bird Survey (BBS) routes? Long-distance migrants are absent during most storms, thus they are a negative control on the same survey. Data: BBS 2025 release, 2,025 US routes east of 100 W and south of 46 N, 65,117 surveys in 1970-2024, 192 species (43 resident, 40 partial migrant, 109 migrant in AVONET), 4.18 million route-species-year counts; Atlantic HURDAT2 tracks 1966-2024. Method: modelled peak wind at each route and season (symmetric modified Rankine vortex, Willoughby-Rahn radius); 497 surveys on 298 routes followed hurricane-force wind (64 kt or more; 75 storms). Event models of log(1+count) with route x species, observer and species x year or species x region x year fixed effects; two placebo seasons after the survey; resident-minus-migrant contrast with route x year fixed effects; two-way clustered intervals; 1,300-draw season permutation; survey-process checks; dose response; four region and period subsets. Results: resident counts were -3.0% (95% CI -5.3 to -0.7) after hurricane-force wind with national year effects and -2.1% (-4.4 to +0.2) with regional year effects; placebo seasons were near zero; partial migrants (+0.4%) and migrants (-0.3%) did not change; estimates two and three seasons later were near zero. The resident-minus-migrant contrast was -3.0% (-5.4 to -0.5) with national and -1.2% (-4.0 to +1.6) with regional year effects. Season permutation: p = 0.055 for residents and p = 0.23 for the contrast. The resident decrease occurred west of 84 W (-4.0%) and in 1998-2024 (-3.3%) but not in the east or in 1970-1997; the contrast did not repeat in any subset. There was no monotonic dose response. Exposure did not change run completion, observer change or survey date. Seven resident woodpecker species decreased by 7.0% (SE 1.6), an unplanned exploratory result. Limitations: simple wind model, no detection model, few exposed surveys clustered in 38 seasons, imperfect migrant control, observational design. Conclusion: any mean resident decrease is small (likely under 4-5%) and not robust.

Breeding Bird Survey · hurricane · tropical cyclone · extreme weather · HURDAT2 · resident birds · migratory birds · negative control · placebo test · permutation test · event study · fixed effects · AVONET · woodpeckers · eastern United States